1. Field of the Invention
This invention relates to electro-optic (EO) modulators and more particularly to hybrid EO polymer/sol-gel modulators in which the sol-gel core waveguide does not lie below the active EO polymer waveguide in the active region.
2. Description of the Related Art
An EO modulator is fundamentally a device that is able to impress an electrical signal on the amplitude or phase of an optical input through the use of special materials that exhibit an EO effect. In such materials, when an electric field is applied to the material, an associated change in refractive index occurs, which can be used to create various kinds of EO modulators, including phase, Mach-Zehnder modulators and directional coupler modulators. The change in refractive index is directly proportional to the applied electric field, so that this is sometimes called the linear EO effect, which only occurs in materials lacking a center of symmetry, as opposed to the much smaller quadratic EO effect which occurs in all materials. In most applications, the EO modulator is part of a fiber-optic communications system, in which case input and output fibers are aligned and attached to an optical waveguide that has been created in the EO material by a variety of methods. EO modulators are generally used when direct modulation of the laser in the communications system is not a viable option, which occurs when very high bandwidth (>10 GHz) or signal linearity are required. High bandwidth is needed in very high bit rate digital communications systems (OC-768 systems running at 40 Gbps), while both high bandwidth and high signal linearity are required for analog applications such as phased array antennas, optical digital to analog converters, and the like. The cw laser is typically a diode laser in the wavelength region from 800-1600 nm, predominantly near 1310 nm (O band) and near 1550 nm (C and L band).
The modulator is characterized by several critical parameters, the most important of which are as follows:                Half-wave voltage, Vπ—the voltage change needed to take the modulator from its maximally transmitting state to its minimally transmitting state; it is generally desired that this voltage be as small as possible;        Insertion loss—the optical loss, in decibels, that is suffered by light as it travels from the input to the output of the modulator; it is always desired that the loss be as small as possible;        Extinction—the ratio of the output power in the “on” state to the output power in the “off” state; this ratio should be as large as possible and is generally measured in decibels; and        3 dB bandwidth—the electrical driving signal frequency at which the maximal optical output of the modulator, driven at the low frequency half-wave voltage, has dropped by 3 dB or 50%.        
For a Mach-Zehnder waveguide EO modulator, the half-wave voltage is related to the various other parameters of interest via:
            V      π        =                  λ        ⁢                                  ⁢        d                              n          eff          3                ⁢                  r          max                ⁢        Γ        ⁢                                  ⁢        L              ,where λ is the optical wavelength, d is the physical separation distance between the drive electrodes in the direction along the applied field, neff is the effective refractive index for light polarized along the direction in which the electric field is applied (generally a function of both waveguide materials properties and dimensions), rmax is the maximum value of the EO coefficient (in units of picometers per volt) that can be achieved in the material, which dictates the preferred directions for the applied field and the light polarization, Γ is the normalized dimensionless overlap integral that measures the degree to which the optical and electrical fields overlap, having a minimum value of 0 and a maximum value of 1 and L is the length of the active region of the modulator, defined as that region where an electric field is applied to an EO material waveguide.
A hybrid EO polymer/sol-gel modulator was first described in Enami, et. al. Appl. Phys. Lett. 82, 490 (2003) and most recently reviewed in Enami, et. al., Nature Photonics 1, 180 (2007) involves using organically modified sol-gels as the cladding and EO polymers as the waveguide in the active region of the modulator and using just sol-gels in the passive region of the modulator where coupling to the optical fiber occurs. These sol-gels have an easily adjustable refractive index, which makes it possible to make waveguides with low coupling loss to optical fiber. They also have been shown to allow for efficient in-device poling of the EO polymer as discussed in C. T. DeRose, et. al. Appl. Phys. Lett. 89, 131102 (2006) and U.S. Pat. No. 7,391,938.
As shown in FIG. 1, a hybrid EO polymer/sol-gel modulator 10 is formed on an insulating layer on a substrate (not shown). The modulator includes a bottom electrode 16, a sol-gel under cladding 18 having a refractive index n1, a sol-gel core 20 having a refractive index n2>n1, and a sol-gel over cladding 22 having a refractive index n1. The sol-gel core 20 is confined below and on either side by under cladding 18. In order to provide a symmetric mode as the light propagates through the waveguide and to provide better coupling to the input and output fibers, the cladding indices n1 are preferably the same. However, some slight variation (<<1%) may occur within a cladding or between cladding layers due to slight variations in fabrication or design.
A vertical taper 23 in layer 22 exposes the surface of core 20 above electrode 16. An EO polymer waveguide 24 having a refractive index n3>n2 covers the exposed surface of the sol-gel core layer. In typical embodiments, the refractive index n3 is designed to be uniform over the active region. Small variations from the uniform value on account of fabrication and poling can occur without degrading performance. A non-uniform index profile may be designed to provide the same accumulated phase change. A buffer layer 26 having a refractive index n4<n2, typically <n1 and preferably <n1−0.1 covers the EO polymer layer. A top electrode 28 defines an active region 30 between itself and bottom electrode 16 and passive regions 32 to either side. A voltage signal Vsig 34 is applied between the top and bottom electrodes to apply an electric field along the poling direction of the EO polymer to change the refractive index of the EO polymer waveguide and modulate the amplitude or phase of light 36 passing through the modulator. Modulator 10 may be configured as a phase modulator 38 having one arm as shown in FIG. 2a or as a Mach-Zehnder modulator 40 having a pair of arms as shown in FIG. 2b and with the ability to produce on-chip amplitude modulation. The Mach-Zehnder can be driven using single-arm, dual-drive or push-pull techniques well known to those in the art.
Light 36 can be input to and output from the modulator using a standard single-mode optical fiber (SMF-28 from Corning); the dimensions of the input and output sol-gel waveguides (˜4 μm×4 μm) and the refractive index difference (˜1%) between the core and the cladding sol-gels can be optimized so that coupling loss to SMF-28 is minimized. Light 36 propagating in sol-gel core 20 proceeds and soon enters the region where the sol-gel over cladding 22 is physically tapered. As the light propagates further into this region it begins to “see” the EO polymer waveguide 24 that has been deposited in the recessed region over and between the physical sol-gel tapers on the input and output sides of the modulator. Since the EO polymer has a significantly higher refractive index (˜1.6-1.7) than the sol-gel core (˜1.5), the light from the sol-gel core is gradually or “adiabatically” pulled up into the EO polymer so that by the end of the taper as much light as possible has been transferred to the EO polymer. The same mechanism (in reverse) results in the transfer of light from the EO polymer back to the sol-gel core at the output. In the active region 30, the refractive index of the buffer layer 26 is often chosen to be very low (˜1.3-1.4) which ensures that the optical waveguide mode in the EO polymer waveguide 24 drops off very rapidly as it enters the buffer layer thereby minimizing losses from the electrodes. In the case of a single-arm Mach-Zehnder modulator (FIG. 2b), when a field is applied to the electrodes and light is propagating in the modulator, light propagating in the left-hand arm of the Mach-Zehnder receives a phase shift relative to the light propagating the right-hand arm, so that the intensity at the output of the Mach-Zehnder changes.
In a representative embodiment sol-gel claddings 18 and 22 and core 20 are comprised of 95/5 (n=1.487) and 85/15 (n=1.50) mixtures (mole %) of methacryloyloxy propyltrimethoxysilane (MAPTMS) and zirconium-IV-n-propoxide, where the MAPTMS provides both organic and inorganic character as well as photopatternability, while the zirconium-IV-n-propoxide is used as an index modifier. The under cladding 18 is deposited by spin coating and hard baked (˜8.5 μm), while the core 85/15 20 is also deposited by spin coating, but then is soft-baked and photopatterned, through a simple wet etching process, prior to hard baking (˜4 μm). Under cladding, also 18, is deposited after the core is patterned to provide confinement on either side of the core. This photopatterned sol-gel waveguide provides excellent coupling to standard SMF-28 single-mode fiber, with coupling losses in the 0.5-1.0 dB per end face range routinely achieved. Propagation losses for this standard sol-gel are 3-4 dB/cm. An adiabatic vertical transition to an EO polymer waveguide (n˜1.65) is accomplished through the use of a grey scale mask that creates a vertical taper in the next layer of sol-gel (i.e. sol-gel over cladding 22). Note that this is not an evanescent field device, as the vertical adiabatic transitions result in more than 70% of the optical field being in the EO polymer waveguide 24. The vertical tapers are also low loss, with less than 1 dB of radiation loss at each taper when the device is made to design. The EO polymer waveguide 24 is typically quite thin (˜1 μm) to achieve single-mode operation; the sol-gel over cladding 22 provides lateral confinement. The buffer layer 26 consists of CYTOP® (Asahi Glass), which has a very low refractive index (n=1.33) at 1550 nm as well as extremely low optical absorption, thereby providing good optical isolation from the top drive electrode 28 even for thin layers (˜1.5 μm).